← 专区目录
目录 · 跳到任一小节
常见生物医学实验 · 第 1 篇 · 一条带怎么变成数字

WB：一条带是怎么变成数字的

划线的词句是中文难以无损传达的地方，鼠标悬停看原文与失味点；点一下就跳到页末注释。
这一页只讲一件事：膜上那条黑线，凭什么能写成“2.3 倍”。答案不在显影仪的屏幕里，而在“这个信号还在不在线性段内”这一句话上。

你迟早会做一次免疫印迹（Western blot, WB），也迟早会听到这句话：“这张膜结果很好，处理组明显比对照组黑。”然后有人把图截下来、圈一圈、算个灰度，写成“表达上调 2.3 倍”。这一页要把这个过程拆开：先弄清 WB 物理上到底在测什么，再走一遍台面流程的关键参数，最后才谈那条黑线怎么变成数字——因为绝大多数“WB 结果不可重复”的争议，都发生在最后这一步，而不是在配胶和上样。

本篇精译自 Bio-protocol 方法学文章 A Simple, Reproducible Procedure for Chemiluminescent Western Blot Quantification（Cañada-García 与 Arévalo，Bio-protocol 2023;13(9):e4667；DOI 10.21769/BioProtoc.4667；PMID 37323629）。原文把台面操作写成编号清单、把缓冲液配方单列一节；中文部分把两者合成一条流程，术语和判读补了教科书级常识，凡是原文没有给出、只能靠经验取值的参数，我都标注为“经验值”。

一、WB 在测什么：三件事叠在一起

把 WB 说成“测蛋白含量”，是错的。它实际输出的是一张灰度的光学图像，而这张图像是三层完全不同的物理过程叠出来的结果。

第一层：按分子量把蛋白摊开。样品在十二烷基硫酸钠-聚丙烯酰胺凝胶电泳（SDS-PAGE）里被处理成“裹上一层负电荷的伸展多肽”。SDS 按大约每克蛋白 1.4 克的比例结合，把不同蛋白的自身电荷差异抹平，于是所有蛋白的荷质比接近一致，泳动速度只由分子量决定。这也是为什么“条带应该在多少 kDa”是一个可核对的硬指标：如果你的目标蛋白预测 55 kDa，条带却出现在 40 kDa，先怀疑抗体不特异，而不是先怀疑生物学发现。

第二层：用抗体把“那一类蛋白”挑出来。转膜之后，膜上其实粘着成千上万种蛋白。一抗（primary antibody）靠可变区识别目标蛋白上的表位（epitope），二抗（secondary antibody）再识别一抗的恒定区并带上辣根过氧化物酶（horseradish peroxidase, HRP）。这一步是特异性问题：抗体认错人，后面测什么都白搭——第 2 篇整篇讲这件事。

第三层：把“抗体结合上了”翻译成“有光”。HRP 催化化学发光底物（chemiluminescent substrate）里的鲁米诺（luminol）在过氧化氢存在下氧化，发出波长约 425 nm 的蓝光。相机（电荷耦合器件，charge-coupled device, CCD）把光积累成像素值。注意“积累”这个词：曝光时间越长，同一个条带的像素值越高，直到某个上限为止。这一句是第 4 节全部讨论的起点。

所以 WB 的信号强度 = （目标蛋白量 × 转膜效率 × 一抗亲和力 × 二抗放大倍数 × 酶活 × 底物浓度 × 曝光时间）的某个非线性函数。式子写出来不是为了让你算，而是为了让你记住：里面每一项都能改，只有第一项是你关心的。这就是原文反复强调它是半定量（semiquantitative）技术的原因——它能回答“两组之间有没有明显差别、方向如何”，不能回答“多少个分子”。

二、“条带黑”为什么不等于“蛋白多”

先看一个能在原文表 1 里直接验证的事实：作者在同一张膜上跑了 6、9、15、18、22.5、30、37.5 µg 六个上样量，用同一套方法算出各自的斜率。结果上样量从 6 µg 到 37.5 µg 只差 6.25 倍，斜率却从 1157.6 涨到 223579.3，差了约 193 倍。如果“黑”真的正比于“多”，这两个倍数应该接近。它们不接近，说明信号链在中后段被放大了，而不是被忠实传递了。原文没有讨论这个偏差，但数字摆在那里，你可以自己看见。

造成这种偏离的机制至少有四条，每一条都值得你在实验记录本上记一笔：

（1）检测器会饱和。CCD 的每个像素有量程上限（常见 16 位图上限 65535），底物也会被耗尽。原文图 2 的图注写得很清楚：37.5 µg 的样品在曝光 828 s 时信号饱和，“红色信号”出现在伪彩图上。饱和之后，再黑也只是“更饱和”，像素值已经不再随蛋白量变化——这时候量出来的任何数字都没有意义。

（2）酶促反应不是线性的。HRP 催化产生的光强依赖底物浓度、pH、温度，以及底物被消耗的速度。膜上局部酶量很高时，底物在几秒内就被耗尽，光强的增长会自动变慢。这就是为什么“曝光 5 s 和曝光 300 s”量出来的比值不一样。

（3）转膜有大小歧视。大分子蛋白从凝胶里转出来更慢、更容易卡在胶里；小分子则可能直接穿过膜。原文给的湿转条件是 200–400 mA、至少 1.5 h、冷房加冰盒，就是为了在“转得动”和“不过热”之间取平衡。如果你的目的蛋白 250 kDa，却和 30 kDa 的内参用同一套转膜时间，前者很可能根本没转干净。

（4）“总量”和“占比”是两件事。上样 20 µg 总蛋白，其中目标蛋白可能只占万分之一。处理组如果总蛋白上样量少了 10%，同时目标蛋白占比翻倍，条带看起来仍然“黑了一倍”，但你归一化之后得到的结论会完全不同。这正是为什么必须有内参，而内参本身也可能变（见第 4 节）。

结论很硬：“条带黑”只在同一张膜、同一批抗体、同一曝光水平、且不饱和的前提下，才能当作相对多少的粗略指标。跨膜、跨批次、跨曝光比较，得到的“倍数”是不可信的。原文的做法就是把这句话变成可操作的程序：不比较单张图，而是比较“信号随曝光增长的斜率”。

三、台面流程与关键参数

下面这条流程照原文的步骤顺序写，参数用原文给的数值；原文没有给出、只能靠经验取的，我写“经验值”并说明理由。每一步后面都加一句“这步错了你会看到什么”——这比记住参数更有用，因为参数可以查，症状只能靠经验认。

① 裂解（lysis）。以 6 cm 皿、80% 融合度的细胞为例，吸干培养液后加 500 µL 冰冷裂解液，冰上或 4 ℃ 轻摇 30 min，收进预冷离心管，16 000×g、4 ℃ 离心 10 min，取上清。原文配方是 20 mM Tris pH 8.0 + 137 mM NaCl + 1% NP-40，用前现加 蛋白酶与磷酸酶抑制剂（protease and phosphatase inhibitors）：1 mM 苯甲基磺酰氟（phenylmethanesulfonyl fluoride, PMSF）、1 µg/mL 抑肽酶（aprotinin）、2 µg/mL 亮抑酶肽（leupeptin）、1 mM 原钒酸钠、10 mM 氟化钠（NaF）、20 mM β-甘油磷酸。原文特别提示：要在冷环境操作，否则蛋白会降解。

这步错了会看到什么：目标条带变弱、下方出现“拖尾”或一串小分子量杂带——那是降解产物；磷酸化蛋白（如 pTrkA）信号忽高忽低，通常是原钒酸钠和 NaF 忘了加，磷酸酶把磷酸基团切掉了。PMSF 在水溶液里半衰期只有几十分钟，必须用前现加，提前配好放着等于没加。

② 定量（quantification）。蛋白浓度必须测，不能“凭感觉上样”。原文明确写了方法要跟裂解液匹配：含 SDS 的裂解液用 BCA 法（bicinchoninic acid），不含 SDS 的用染料法（Bradford 类试剂）。原因是去垢剂、还原剂会干扰显色反应。BCA 的标称线性范围常写作 20–2 000 µg/mL（这是试剂盒说明书上的典型值，具体以你实验室那盒的说明书为准）；低于下限时信号接近空白，高于上限时标准曲线会弯。做标准曲线要复孔，样品通常稀释 5–20 倍后落进曲线中段。

这步错了会看到什么：所有泳道的条带“整齐地一样黑”，或者内参与目标蛋白一起同向变化——多半是上样量本身算错了；标准曲线 R² 很高但样品读数落在曲线之外，则是最常见的隐性错误：数字很漂亮，值不可信。

③ 上样与配胶。取 15–50 µg 蛋白（原文的经验区间），加 上样缓冲液（Laemmli buffer）到 2× 终浓度，补足体积到 25 µL，混匀后 100 ℃ 煮 6 min，室温冷却，16 000×g 离心 30 s；不马上跑可存 −20 ℃。胶的浓度按分子量选：大于 80 kDa 用 6%–8%，小于 30 kDa 用高浓度胶；若同一张胶上既有小于 30 kDa 又有大分子，用 4%–12% 梯度胶。原文的胶配方是 350 mM Bis-Tris pH 6.4 + 0.1% SDS + 0.1% 过硫酸铵（APS）+ 6.6 mM 四甲基乙二胺（TEMED），分离胶 4.5 mL、浓缩胶 2 mL，上层用异丙醇封液面隔绝氧气；自配胶 4 ℃ 湿盒可存 3–4 天。

这步错了会看到什么：条带整体“歪”或呈微笑状——胶没凝匀或电压过高发热；泳道里有白色不溶物——样品没煮透或离心不彻底；所有条带都比预期低 5–10 kDa——胶的 pH 或丙烯酰胺比例不对，也可能是 marker 用错了。

④ 电泳（SDS-PAGE）。原文给的两组条件是：Bis-Tris 胶 200 V、30–40 min；Tris-Glycine 胶 90 V、75–120 min。判据不是时间，而是溴酚蓝前沿跑到胶底部。大分子建议降低电压、延长时长，让它们有时间“挤”进胶里。

这步错了会看到什么：前沿跑出胶外、小分子条带丢失；或跑太快导致条带变宽、分辨率下降（两条靠得很近的带糊成一条）。

⑤ 转膜（transfer）。用 0.2 µm 的聚偏二氟乙烯膜（polyvinylidene difluoride, PVDF）。PVDF 是疏水的，必须先活化：甲醇浸数秒 → 双蒸水洗两次、每次 3–4 min → 泡进转膜液备用。转膜液配方是 25 mM Tris pH 8.3 + 192 mM 甘氨酸 + 10% 甲醇，4 ℃ 保存、最多重复用 5 次。湿转条件：200–400 mA、至少 1.5 h、冷房中进行并加冰盒与搅拌子；原文也说明半干转（semi-dry transfer）可以用，只是参数不同。夹层顺序必须是“胶在负极侧、膜在正极侧”，装好后用滚轮赶气泡。

这步错了会看到什么：整张膜几乎没信号而胶上还留着彩色预染 marker——方向装反了，蛋白跑到缓冲液里去了；局部白斑——气泡；大分子完全不出——转膜时间/电流不够；膜上出现“糊成一片”的低分子量背景——转过了头（小分子穿透膜）。

⑥ 封闭与抗体孵育。用封闭液（blocking solution）室温摇至少 30 min。原文给了两种配方：一抗稀释用 TBST + 3% 牛血清白蛋白（bovine serum albumin, BSA）+ 0.05% 叠氮钠，二抗用 TBST + 5% 脱脂奶粉、用前现配（摇动时避免起泡）。一抗孵育：室温 2–3 h 或 4 ℃ 过夜；二抗是 HRP 标记的羊抗鼠/羊抗兔，工作浓度 1:5 000，孵育至少 45 min。洗膜用 TBST（10 mM Tris pH 7.5 + 150 mM NaCl + 1 mM 乙二胺四乙酸（EDTA）+ 0.1% 吐温-20），至少洗 3 次、每次至少 15 min。

这步错了会看到什么：整张膜均匀发灰——封闭不足或二抗浓度过高；只有泳道位置有信号、其余地方全是黑背景——膜干了；背景里出现一堆不认识的小带——一抗不特异（第 2 篇的主题）；信号特别弱而阳性对照也弱——抗体反复冻融失活，或叠氮钠污染了 HRP 体系（叠氮钠抑制 HRP，不能加到二抗稀释液里）。

⑦ 显影与曝光。膜上滴 增强型化学发光底物（enhanced chemiluminescence, ECL），原文是孵育 1 min 后进暗箱拍照。原文的关键操作是：不要只拍一张。相机软件里把曝光间隔设成 1、3、5、15、30、60、180、300、600、900 s，而实际记录的是累积时间——也就是 1、3、8、18、33、63、108、168、228、288、408、588、828、1 128、1 428、1 728、2 028、2 628、3 228、4 128 s。最后再开白光拍一张带 marker 的膜，方便回头认条带位置。

这步错了会看到什么：第一张就已经全白/全黑——底物加太多或曝光设置错；条带边缘“发胖”、中心变平——已经饱和，见第 4 节；同一批膜里有的条带清楚、有的糊——ECL 分布不均，需要把膜摊平、液体铺满。

四、把条带变成数字

原文这篇方法学文章的全部价值，浓缩成一句话：不要比单张图，要比“信号随曝光增长的速度”。这一节按它的实际步骤走，并补上它没有明说、但你必须知道的边界条件。

（1）先确定信号还活着的那一段。把同一张膜的一串曝光图拼起来看，你会看到两条曲线的形状：曝光短的时候信号随曝光时间近似直线上升，这叫线性范围（linear range）；曝光再长，曲线开始弯、最后拉平——那是饱和。原文图 2 的图注给了具体例子：37.5 µg 上样量在 828 s 时饱和。所以第一步不是量条带，而是找出所有样品共同的线性区间：在这段曝光时间里，每个样品的信号都还没弯。原文示例里作者手动选的区间是 63–828 s（共 12 张图），选完之后所有样品都用同一段区间拟合。

为什么必须“共同”？因为如果 A 组在 100 s 就饱和、B 组在 800 s 才饱和，你从某一单张图里量出来的 A/B 比值，会随着你挑哪张图而变——挑得巧就是 2.0，挑得巧就是 1.3。对饱和条带做定量，得到的数字不是“偏大”或“偏小”，而是没有定义。

（2）用 ImageJ 逐张量，矩形大小不许变。操作链是：File → Import → Image Sequence，选中文件夹里所有化学发光图（不要把那张白光拍的 marker 图导入）；Analyze → Set Measurements 勾选 Area、Mean gray value、Integrated density（积分密度）与 Display label；用矩形工具框住一条带，然后把同一个矩形平移到条带旁边没有信号的膜区域量一次背景；再用 Image → Stacks → Measure Stack 把这个矩形在整组曝光图上逐张量完。每一条带重复这个动作，最后存成 CSV。

原文特意强调两点：矩形要尽量贴紧条带（框得越大，背景引入的误差越大），以及矩形大小必须固定——因为一旦你“手工把框拉大一点”，信号强弱就会混进面积里，不同样品之间比的就不再是同一个几何区域。它还建议：如果胶上有空泳道，就上一孔“只有裂解液 + 上样缓冲液”的空白样，专门用来量背景。

（3）扣背景，再拟一条直线。把每个条带在每个曝光下的积分密度减去同一张图上、同一曝光下的背景值，得到“净信号”。以曝光时间（累积秒数，不是相机软件里那个间隔值）为横轴、净信号为纵轴画散点，你会看到每条带各自是一条上升的直线。对每条带在线性区间内做简单线性回归，取斜率作为这条带的代表值。原文表 1 给的就是这套结果：R² 在 0.9987–0.9998 之间，p 值都在 10⁻¹⁰ 量级；六个上样量的斜率依次是 1 157.6、3 156.1、16 760.6、26 829.7、57 338.1、121 563.9、223 579.3（单位为“积分密度/秒”，依次对应 6、9、15、18、22.5、30、37.5 µg）。

判断拟合好不好，看两件事：R² 是否接近 1（原文的判据是“不够接近 1 就回去重新选曝光区间”），以及散点图有没有明显弯折。这两条都不满足时，正确的反应是回去重选区间，而不是“把不好看的那几个点删掉”。

（4）内参：它也是会变的。斜率给的仍然是“这条带的信号强度”，不是“目标蛋白的绝对量”。要比较不同样品，必须消除上样量的差异，这就是内参（housekeeping protein 或 reference protein）的用途。原文的表述是：用“不随实验条件变化”的蛋白来归一化，通常是肌动蛋白（actin）、微管蛋白（tubulin）、甘油醛-3-磷酸脱氢酶（GAPDH）。

注意这句话里的循环：“不随实验条件变化”是内参的定义，不是它的已知性质。GAPDH 是糖酵解酶，缺氧、缺血、氧化应激都会上调它；actin 在细胞骨架重塑、凋亡小体形成时会变；tubulin 在轴突损伤、微管解聚的模型里也会变。你做的如果是脑出血、缺血再灌注这类模型，这三条都踩在“会变”的名单上。

可操作的做法（经验值，但都有依据）：在同一处理条件下先单独跑一次内参，看各组内参条带的斜率差异是否在 ±20% 以内；超了就不能用它，换一个（如 vinculin、HSP90、β-tubulin 视组织而定），或者做两个内参取几何平均；更稳妥的是加一个与抗体无关的上样对照——丽春红（Ponceau S）染色或免染总蛋白（stain-free）成像，直接看每条泳道的总蛋白量。

（5）归一化，然后才谈倍数。每个样品的相对水平 = 目标蛋白斜率 ÷ 同一泳道内参斜率；再把各组的这个比值除以对照组的均值，得到归一化（normalization）后的相对倍数。原文的示例就是这么做的：先算磷酸化 TrkA（pTrkA）与总 TrkA 的比值，再以刺激 5 min 那个时间点为 1 做归一化，最后画三次独立实验的结果。

要清楚归一化做不到什么：它消不掉抗体批间差异、消不掉跨膜差异（原文因此强调对照与实验样品必须放同一张膜）、也消不掉饱和带来的非线性。它只处理“上样量不一致”这一件事。

（6）重复：n 是生物学重复，不是胶片张数。同一管裂解液跑三块胶，是技术重复，只能告诉你“手法稳不稳”；三次独立的细胞/动物实验才是生物学重复。原文示例的 n = 3，图注写明“每条彩色线代表一次独立实验”。统计检验应当建立在归一化后的比值上，而不是拿未归一化的灰度值直接做 t 检验——后者把上样量误差算成了生物学差异。

五、一张膜该怎么读

下面这张对照表不是评分标准，而是“先看哪里”的顺序。你在组会上被问“这张 WB 能信吗”，就按这个顺序答。

看什么可以接受需要追问
条带数量目标位置单一条带，或在预期分子量处有明确主带多条带却不解释；主带位置与预测差得远（先怀疑抗体）
分子量标注图上标了 marker 的 kDa 数，能对得上整张图没有 marker，或 marker 被裁掉
内参各组内参条带粗细基本一致，且内参本身有线性内参与目标蛋白同向大幅变化；不同图的内参带长得一模一样
饱和作者报告了曝光序列，说明了线性区间只给一张“最好看”的图，没有任何曝光信息
重复n ≥ 3，来自独立实验，图注写明“n = 3”其实是同一批样品跑三次
原始图能提供未裁剪的全膜图只给裁剪过的条带；相邻条带边缘对不齐

最后一行值得单独说：裁剪本身不是造假，裁剪是为了排版；问题在于裁剪后无法核对。学界现在普遍要求投稿时提供未裁剪的原始膜图，理由很朴素——条带的形状、背景的纹理、相邻泳道的间距，都是判断“这张图有没有被拼过”的线索。

六、本科生一定会踩的坑

① 按体积上样，不按蛋白量。“每个孔上 10 µL”听着整齐，其实是在赌所有样品浓度一样。正确做法是先定量、再算每个样品该取多少 µL、用裂解液补到同一体积。

② 煮沸后不离心。煮完的样品会有不溶物，直接上样会堵泳道、把条带顶歪。原文明确写了煮后 16 000×g 离心 30 s。

③ 胶没凝就拔梳子。梳子拔早了泳道会塌，拔晚了两侧会撕裂。看胶面是否还有流动的液面、以及剩余胶液是否已凝固，比看时间可靠。

④ 转膜方向装反。这是新手最经典的一次性报废：整张膜干干净净，胶上还留着蓝色 marker。

⑤ 洗膜不换液、不摇匀。背景高、出现“地图状”黑斑，多半是洗液没有完全覆盖膜、或者洗的时间不够。原文的要求是至少 3 次、每次至少 15 min。

⑥ 一抗反复冻融。抗体里的蛋白会聚集，信号越来越弱、背景越来越高。分装成小份、一次一管，是最省钱的操作。

⑦ 一上来就长曝光。先拍一张 5 s 看有没有信号，再从短到长拍序列；反过来做，等你发现饱和时，底物已经消耗掉一部分了。

⑧ 拿“看起来黑”当结论。这是本篇要拆掉的那个念头。条带黑只说明“在这个曝光下这个位置亮”，要说“蛋白多”，得先证明信号在线性段内、并且被内参归一化过。

七、以你现在的水平能亲手做到哪一步

这一节不是给你打气，是给你一张“权限表”。判断标准只有两条：这一步做错的代价有多大，以及错了以后能不能看出来。代价小、错了立刻能看见的，你现在就该自己上手；代价大、错了要等到一周后才发现没信号的，必须在有人看着的时候做。

可以自己来（错了能返工，且症状明显）：

① 配胶。照着配方算体积、按顺序加 APS 和 TEMED、灌分离胶、用异丙醇封液面、凝固后灌浓缩胶、插梳子。配坏了最多重配一次，成本几块钱。唯一要养成习惯的是：APS 和 TEMED 最后加，加完立刻灌，动作慢了就凝在烧杯里。

② 上样。在师姐给你算好“每个样品取多少 µL、补多少缓冲液”之后，你自己加样。加样枪别插到底、别吹泡、别把样品带进相邻泳道——这三件事只有靠手熟。

③ 跑胶与洗膜。接电源、设电压、看前沿、按时间换洗液。这些都是“看得见”的操作。

④ 配溶液。TBST、转膜液、封闭液，按配方称量、调 pH、定容。这一步最容易出低级错误（称错、pH 没调、忘了加吐温-20），也最适合新手练手，因为错了当天就能发现。

⑤ ImageJ 量灰度。用固定大小的矩形、逐张量、扣背景、存 CSV。这个操作本身不难，难的是判断“哪些图该量”——那属于下一档。你可以先量，然后把结果交给师姐核对。

必须在师姐看着时做（错了代价大或不可逆）：

⑥ 裂解。抑制剂必须用前现加、全程冰上、离心机预冷到 4 ℃。任何一步疏忽，蛋白就降解了，而你当天看不出来——要等到曝光时才发现目标条带消失。这一步的失败成本是一整批样品。

⑦ BCA 定量。标准曲线怎么稀释、样品稀释多少倍、复孔怎么排、显色时间到了没有，都会把浓度算错。浓度错了，后面所有“上样量一致”的假设都塌了，而你在膜上是看不出来的。

⑧ 一抗稀释与孵育。一支抗体几百到几千块，稀释错了就浪费；浓度过高会带来满屏背景，过低则一片空白。第一次做，请让师姐在旁边看着你算稀释倍数、看着你把抗体加进孵育盒。

⑨ 转膜装配与电流设定。夹层顺序、赶气泡、冰盒、搅拌子、电流大小，任何一项出错都是整张膜报废。这一步是我最不建议你“自己先试试”的。

⑩ ECL 加量与曝光判断。底物加多少、什么时候开始拍、拍到哪一张算饱和——这需要有人指着屏幕告诉你“这张已经平了”。你自己拍，很可能拍到一张“特别好看”的饱和图，然后把它当成结果。

现在还轮不到你独立判断（需要经验，也需要对整个实验负责）：

⑪ 这个模型该用哪个内参；⑫ n 要做几次才算够；⑬ 两组之间那 1.3 倍的差别算不算“有差异”；⑭ 这张膜是重做还是可以用；⑮ 这条带到底是不是你的目标蛋白（这需要结合分子量、阳性对照和抗体说明书一起判断）。

最后一句提醒：你第一次做 WB 的目标不是“做出好看的图”，而是把整条流程走完、并留下能被复核的记录——哪天裂解的、上样量多少、抗体是哪一批、拍了哪些曝光时间。这些记录在你三个月后回看时，比那张图值钱得多。原文之所以能被称为“可重复”，靠的也是这个：它把每一步的参数都写死了，包括那一串间隔与累积曝光时间。

八、参数速查表

下表把原文给出的关键参数集中在一处，方便你在实验台前对着看。标“经验值”的是原文没有、只能按常规取的；标“说明书”的以你手上那盒试剂为准。

步骤关键参数（原文值）这一步错了的症状
裂解6 cm 皿加 500 µL 冰冷裂解液；4 ℃ 30 min；16 000×g、4 ℃、10 min；抑制剂用前现加条带变弱、下方拖尾（降解）；磷酸化信号忽高忽低
定量含 SDS 用 BCA，不含 SDS 用染料法；线性范围约 20–2 000 µg/mL（说明书）各泳道条带“整齐一致”；浓度读数落在曲线外
上样15–50 µg；2× Laemmli；补足至 25 µL；100 ℃ 6 min；16 000×g 30 s泳道堵塞、条带歪斜；目标条带整体偏弱
配胶>80 kDa 用 6%–8%；<30 kDa 用高浓度；跨范围用 4%–12% 梯度胶大分子挤不进胶、小分子跑出胶
电泳Bis-Tris：200 V、30–40 min；Tris-Glycine：90 V、75–120 min；跑到前沿到底条带变宽、相邻带糊在一起
转膜0.2 µm PVDF，甲醇数秒 → 水洗 2×3–4 min；25 mM Tris pH 8.3 + 192 mM 甘氨酸 + 10% 甲醇；200–400 mA、≥1.5 h、冷房整膜无信号（方向反）；局部白斑（气泡）；大分子缺失（时间不足）
封闭室温摇 ≥30 min；一抗用 3% BSA + 0.05% 叠氮钠；二抗用 5% 脱脂奶粉现配整膜均匀发灰；背景高
一抗室温 2–3 h 或 4 ℃ 过夜（稀释比按说明书）空白或满屏杂带
二抗HRP 标记，1:5 000，≥45 min信号过强/背景高；或几乎无信号
洗膜TBST，≥3 次、每次 ≥15 min背景斑块、条带边缘发毛
曝光间隔 1、3、5、15、30、60、180、300、600、900 s；累积 1–4 128 s；37.5 µg 于 828 s 饱和条带中心发平、边缘发胖（饱和）
定量ImageJ 积分密度；矩形固定；逐张量；扣同图背景；线性区间内拟合斜率；R² 接近 1“倍数”随挑哪张图而变

最后把这一页压成一句：WB 给你的不是“蛋白有多少”，而是“在某个线性窗口里，这条带的信号增长得比那条快多少”。原文的全部设计——拍一串曝光、取共同区间、比斜率、再除以内参——都是为了让这句话里的每个条件都成立。你要做的，就是在自己的实验里把这些条件一条条守住。

▲ 本页翻译难点（对照学习用）

〔注1〕
原文
a universally used technique to identify specific proteins from a heterogeneous and complex mixture
直译
一种被普遍使用的技术，用于从异质而复杂的混合物中鉴定特定蛋白。
失味点
identify 是“认出是谁”，不是“数出有多少”。原文全篇在讲定量，但它对技术定位的用词始终停在 identify——这不是谦虚，是准确。
学习提示
看到任何把 WB 写成“测定蛋白含量”的表述，先在心里打个问号：它越过的时间尺度与线性范围，靠什么补回来？
〔注2〕
原文
proteins are electrophoretically separated in an acrylamide gel according to their size（原文 Procedure 节）
直译
蛋白在丙烯酰胺凝胶中按其大小被电泳分离开。
失味点
according to their size 是 SDS 的功劳，不是蛋白自己的性质：SDS 把蛋白变成带负电的棒状物，才让“大小”成为唯一变量。省掉 SDS 这三个字母，整句话就不成立了。
学习提示
“按分子量分离”是一句承诺：条带的位置必须能被预测。位置不对时，它是你手里最便宜的抗体质检工具。
〔注3〕
原文
antigenic epitopes displayed by the target protein（原文 Procedure 节）
直译
目标蛋白所展示的抗原性表位。
失味点
displayed（展示）不是修辞：WB 里蛋白被 SDS 变性、被固定到膜上，它“展示”的是线性的一段氨基酸序列，而不是天然构象。所以能在 WB 里用的抗体，未必能在免疫荧光里用——那个场景看的是折叠后的三维结构。
学习提示
“这个抗体 WB 能用”不等于“这个抗体什么都能用”。这是第 2 篇的核心，也是实验室里最常见的越界引用。
〔注4〕
原文
the increase in chemiluminescent signal；Chemiluminescent substrate
直译
化学发光信号的增长；化学发光底物。
失味点
“增长”是这篇方法的关键词。它测的不是某一刻的亮度，而是亮度随曝光增长的速度——把静态的“亮”换成动态的“斜率”，饱和问题才被绕开。
学习提示
以后看到任何发光/荧光定量，先问一句：我量的是亮度，还是亮度的增长率？
〔注5〕
原文
Western blotting is a semiquantitative technique（原文 Useful tips 节）
直译
免疫印迹是一种半定量技术。
失味点
semi- 在这里不是“做了一半”，而是“只保证序关系”：能可靠地说 A > B，但说不出 A 是 B 的几倍。中文写“半定量”容易被读成“不太准的定量”，其实它是一句责任声明。
学习提示
凡是要写“2.3 倍”的地方，都得先回答：这个倍数在几条独立实验里稳定吗？换一张膜还在吗？
〔注6〕
原文
To avoid protein degradation, include protease inhibitors and work in cold conditions, at 4 °C.
直译
为避免蛋白降解，需加入蛋白酶抑制剂，并在 4 ℃ 的冷环境中操作。
失味点
这句是一条因果链的两端：抑制剂（化学）＋低温（物理），两个都要。只加抑制剂不冰上，酶活被压低但没消失；只冰上不加抑制剂，低温只是让降解变慢。
学习提示
PMSF 在水里几十分钟就水解，所谓“用前现加”不是仪式，是化学事实。
〔注7〕
原文
5× Laemmli buffer
直译
5× 莱姆利上样缓冲液。
失味点
Laemmli 是人名（U. K. Laemmli），不是成分。它的作用有三件：SDS 继续变性、β-巯基乙醇/二硫苏糖醇（DTT）打断二硫键、甘油让样品沉进泳道、溴酚蓝给你一条看得见的前沿。
学习提示
非还原性 WB（不加还原剂）会看到二硫键连接的二聚体，分子量翻倍——那不是杂带，是你自己没还原。
〔注8〕
原文
Blocking solution will depend on the primary antibody to be used. Check manufacturer datasheet for that info.
直译
封闭液取决于所用的一抗，请查看厂商说明书。
失味点
“取决于”背后是两条经验：脱脂奶粉含酪蛋白（磷酸蛋白），会干扰磷酸化抗体的信号；BSA 不含酪蛋白但可能含微量 IgG，会与某些二抗交叉反应。所以不是随便挑一种。
学习提示
做磷酸化蛋白时用 5% BSA 而不是奶粉，是实验室里的默认动作，理由就在这一句。
〔注9〕
原文
HRP-conjugated anti-mouse antibody (concentration used: 1:5,000)
直译
HRP 标记的抗小鼠抗体（使用浓度为 1:5 000）。
失味点
1:5 000 是稀释比，不是浓度。稀释比相同的两支抗体，实际工作浓度可能差十倍——这正是 2026 年那篇抗体验证共识要求同时报告“稀释比与蛋白浓度”的原因（见第 2 篇注 4）。
学习提示
记实验记录时，两个数都写下：你配的稀释比，以及瓶子上标的原始浓度。
〔注10〕
原文
the slope of the signal increase within the combined linear range of detection
直译
在检测的共同线性范围之内，信号增长的斜率。
失味点
combined（共同的）是整篇方法的技术核心，也是最容易被漏译的一个词：区间必须对所有样品同时成立，而不是每个样品各取一段。少了它，方法就退化成“挑一张好看的图”。
学习提示
写方法时请把区间写出来（例：63–828 s，共 12 张图），不要只写“在线性范围内”。
〔注11〕
原文
housekeeping proteins including actin, tubulin, and glyceraldehyde 3-phosphate dehydrogenase
直译
管家蛋白，包括肌动蛋白、微管蛋白和甘油醛-3-磷酸脱氢酶。
失味点
housekeeping（管家）是个承诺：假设它们在所有条件下都稳定。这个承诺在代谢、缺氧、细胞骨架重塑的模型里经常不成立，而中文“管家蛋白”四个字读起来像个保证。
学习提示
先验证再使用：在同一处理条件下单独跑一次内参，各组差异超过 ±20%（经验阈值）就换。
〔注12〕
原文
it is required to perform data normalization using proteins that do not vary due to experimental conditions
直译
需要用不随实验条件变化的蛋白来做数据归一化。
失味点
这是定义式的循环：先要求“不变”，再用它去除变异。原文把它当常识写过去，但真正的操作是——你得先证明它不变，才能用它归一化。
学习提示
归一化只解决“上样量不一致”。它解决不了抗体批间差、跨膜差、饱和。别指望一个除法能修好四件事。
〔注13〕
原文
Routinely, we set time differences between exposures … of 1, 3, 5, 15, 30, 60, 180, 300, 600, and 900 s. The signal of each image corresponds to the accumulated time from the beginning
直译
常规设置曝光间隔为 1、3、5、15、30、60、180、300、600 与 900 s；每张图的信号对应从开始累加的时间。
失味点
interval（间隔）与 accumulated（累积）是两套数：软件里设的是间隔，做回归时要用累积。把间隔当累积代入，斜率会整体算错，而 R² 可能依然漂亮——错误不会自己报错。
学习提示
原文示例的累积序列是 1、3、8、18、33、63、108、168、228、288、408、588、828…s。抄参数时把这个序列一起抄下来。
〔注14〕
原文
Integrated density（ImageJ 的 Set Measurements 项）
直译
积分密度。
失味点
它等于“选区面积 × 平均灰度”，所以框得越大数值越大。这意味着它不是一个客观量，而是你画的那个框的属性——所以原文要求矩形大小固定，并建议在空泳道上量背景。
学习提示
跟 Mean gray value 比，积分密度对大面积弱带更敏感；跟 RawIntDen 比，它已经扣掉了“测量设置”的影响。选哪个要写进方法，不要中途换。

附录 · 英文原文

⚠ 以下为原文（摘要与背景部分）的忠实转录，仅作术语与句式对照，供你核对中文部分里的引文。原文的完整方法、配方与步骤清单请见 DOI：10.21769/BioProtoc.4667（开放获取，CC BY-NC）。

A Simple, Reproducible Procedure for Chemiluminescent Western Blot Quantification

Journal: 2023;13(9):ee4667  |  DOI: 10.21769/BioProtoc.4667  |  PMID: 37323629  |  PMCID:

Authors: Daniel Cañada-García; Juan C. Arévalo

Abstract

Western blotting is a universally used technique to identify specific proteins from a heterogeneous and complex mixture. However, there is no clear and common procedure to quantify the results obtained, resulting in variations due to the different software and protocols used in each laboratory. Here, we have developed a procedure based on the increase in chemiluminescent signal to obtain a representative value for each band to be quantified. Images were processed with ImageJ and subsequently compared using R software. The result is a linear regression model in which we use the slope of the signal increase within the combined linear range of detection to compare between samples. This approach allows to quantify and compare protein levels from different conditions in a simple and reproducible way.

Graphical overview

Background

Proteins are key biomolecules for the correct functioning of living animals; therefore, it is required not only to detect but also to accurately quantify them. Western blot is the most widely used and accepted method to detect proteins of interest using specific antibodies. Since the development of western blotting in 1979 ([Towbin, H, PNAS 1979]), more than 382,000 articles have been published using this technique (Western blot, PubMed, October 11, 2022). Since 2000, when 7,300 papers were published, there has been a constant increase in the number of articles reporting the use of this method ([Moritz, 2020]), reaching almost 22,600 in 2021. However, a proper quantification to determine the amount or the corresponding protein modification detected is still missing in most of the articles. In addition, reviewers and editors are eager to have trusty quantitative data regarding western blot analyses ([Murphy and Lamb, 2013]; [Taylor et al., 2013;] [Janes, 2015]; [McDonough et al., 2015;] [Bass et al., 2017;] [Pillai-Kastoori et al., 2020]). To assess accurate quantification, it is required to perform data normalization using proteins that do not vary due to experimental conditions. Usually, these are housekeeping proteins including actin, tubulin, and glyceraldehyde 3-phosphate dehydrogenase. Previous articles quantifying chemiluminescence western blots have relied on the use of single images of signals from different samples that are sometimes saturated, leading to incorrect outcomes. Fluorescence-based techniques using fluorophore-conjugated secondary antibodies are not commonly employed in most laboratories but provide a wider range of linear correlation between signal and protein abundance ([Huang et al., 2019]). Therefore, it is urgent to establish a simple, reliable method to quantify western blot signals to draw accurate conclusions in the studies.

Here, we describe an improved, simple protocol for western blot quantification taking advantage of chemiluminescent detection by ccd cameras and ImageJ and R software. Different images with different exposure times were utilized to generate the linear range for each protein to be quantified. The advantage of our procedure is that the signals are quantified across an interval of different exposures within the linear range for all the samples, where the differences between them are constant. Comparison of the slopes obtained provides a more accurate representation of the relative amounts of proteins in different samples than would be achieved by comparing individual images.

Materials and Reagents

The following list provides examples of the materials and equipment that we routinely use in our laboratory. Reagents and equipment with similar specifications will work as well.

All the materials and reagents listed were handled and stored following manufacturer’s instructions.

  1. Polyvinylidene difluoride transfer membranes, 0.2 µm (PVDF membranes) (Thermo Fisher Scientific, catalog number: 88520)
  2. Whatman 3MM CHR paper (Cytiva, catalog number: 3030-335)
  3. β-glycerophosphate (AppliChem, catalog number: A2253, CAS number: 819-83-0)
  4. β-mercaptoethanol (VWR BHD Chemicals, catalog number: 25351.182, CAS number: 60-24-2)
  5. Acrylamide-bis solution (BioAcryl-P 30%, 37.5:1) (Alfa Aesar, catalog number: J61505)
  6. Ammonium persulfate (APS) (Sigma-Aldrich, catalog number: A3678, CAS number: 7727-54-0)
  7. Aprotinin (Sigma-Aldrich, catalog number: A1153, CAS Number: 9087-70-1)
  8. BCA protein assay: reagent A (Thermo, catalog number: 23228) and reagent B (Thermo, catalog number: 23224)
  9. Bis-Tris ultrapure (Thermo Fisher Scientific, catalog number: J12112.22, CAS number: 6976-37-0)
  10. Bovine serum albumin (BSA) (Sigma-Aldrich, catalog number: A7906, CAS number: 9048-46-8)
  11. Bromophenol blue (Amresco, catalog number: 0449-25G, CAS number: 115-39-9)
  12. Dye reagent concentrate (Bio-Rad, catalog number: 500-0006)
  13. Ethylenediaminetetraacetic acid (EDTA) (VWR Life Science, BHD Prolabo Chemicals, catalog number: 443885J, CAS number: 6381-92-6)
  14. Glycerol (Thermo Fisher Scientific, catalog number: BP229-1, CAS number: 56-81-5)
  15. Glycine (Sigma-Aldrich, catalog number: G8898-1KG, CAS number: 56-40-6)
  16. HRP-conjugated anti-mouse antibody (Jackson ImmunoResearch, catalog number: 115-035-003, concentration used: 1:5,000)
  17. HRP-conjugated anti-rabbit antibody (Jackson ImmunoResearch, catalog number: 111-035-003, concentration used: 1:5,000)
  18. Hydrochloric acid (HCl) (Thermo Fisher Scientific, catalog number: H/1200/PB15, CAS number: 7647-01-0)
  19. Isopropanol (Sigma-Aldrich, catalog number: 1.09634.1011, CAS number: 67-63-0)
  20. Leupeptin (Sigma-Aldrich, catalog number: L9783, CAS number: 24125-16-4)
  21. MES-SDS running buffer (Thermo Fisher Scientific, catalog number: NP0002)
  22. Methanol (Fisher Chemical, catalog number: M/4000/PB17, CAS number: 67-56-1)
  23. N,N,N',N'-tetramethyl-ethane-1,2-diamine (TEMED) (Thermo Fisher Scientific, catalog number: BP150-20, CAS number: 110-18-9)
  24. Non-fat dry milk (Nestlé Sveltesse)
  25. NP-40 (IGEPAL CA-630) (Sigma-Aldrich, catalog number: I3021, CAS number: 9002-93-1)
  26. Phenylmethanesulfonyl fluoride (PMSF) (CalbioChem, catalog number: 52332, CAS number: 329-98-6)
  27. Sodium azide (MERK, catalog number: 1066880100, CAS number: 26628-22-8)
  28. Sodium chloride (NaCl) (Sigma-Aldrich, catalog number: 71383-5KG, CAS number: 7647-14-5)
  29. Sodium dodecyl sulfate (SDS) (Bio-Rad, catalog number: 1610302, CAS number:151-21-3)
  30. Sodium fluoride (NaF) (Sigma-Aldrich, catalog number: S7920-100G, CAS number: 7681-49-4)
  31. Sodium orthovanadate (Sigma-Aldrich, catalog number: S6508, CAS number: 13721-39-6)
  32. Trizma base (Tris) (Sigma-Aldrich, catalog number: T1503-1KG, CAS number: 77-86-1)
  33. Tween-20 (Sigma-Aldrich, catalog number: P5927, CAS number: 9005-64-5)
  34. WesternBrightTMECL HRP Substrate (Advansta, catalog number: K-12045-D50)
  35. Lysis buffer (see Recipes)
  36. 5× Laemmli buffer (see Recipes)
  37. Polyacrylamide gels (see Recipes)
  38. Transfer buffer (see Recipes)
  39. TBST washing buffer (see Recipes)
  40. Primary antibody blocking solution (see Recipes)
  41. Secondary antibody blocking solution (see Recipes)

Equipment

  1. Mini-PROTEAN®Tetra Vertical Electrophoresis Cell, 4-gel, for 1.0 mm thick handcast gels (Bio-Rad Laboratories Inc., catalog number: 1658001FC)
    • Mini-PROTEAN Tetra Companion Running Module (Bio-Rad Laboratories Inc., catalog number: 1658038)
    • Casting frame (Bio-Rad Laboratories Inc., catalog number: 1653304)
    • Casting stand gaskets (Bio-Rad Laboratories Inc., catalog number: 1653305)
    • Glass plates (Bio-Rad Laboratories Inc., catalog number: 1653311)
    • Short plates (Bio-Rad Laboratories Inc., catalog number: 1653308)
    • Combs, 15-well, 1.0 mm (Bio-Rad Laboratories Inc., catalog number: 1653360)
    • Buffer tank and lid with power cables (Bio-Rad Laboratories Inc., catalog number: 1658040)
    • Gel releasers (Bio-Rad Laboratories Inc., catalog number: 1653320)
    • Blot roller (Bio-Rad Laboratories Inc.)
  2. Power supply (VWR, catalog number: HITAK622113)
  3. MicroChemi System (Bio-Imaging Systems)

Software

  1. ImageJ and R software

Procedure

The composition of solutions and buffers used are shown below under the section “Recipes”.

  • Western blotSince this protocol aims to quantify the signal from membranes, we provide here a general western blotting guideline, based on which each user should apply the optimal conditions for his or her interest. In brief, proteins are electrophoretically separated in an acrylamide gel according to their size and then transferred to a membrane, to be probed with antibodies that recognize specifically antigenic epitopes displayed by the target protein.The steps are the following:
    1. Extract proteins from HEK293FT cells (at 80% confluency) seeded in a 6 cm plate adding 500 µL of cold lysis buffer with protease inhibitors and incubate for 30 min at 4 °C with gentle shaking. Collect the solution with the cells in a cold Eppendorf tube and centrifuge at 16,000× gfor 10 min at 4 °C.Note: To avoid protein degradation, include protease inhibitors and work in cold conditions, at 4 °C. Lysis can be performed using different lysis buffers including RIPA if e.g., immunoprecipitation will be performed.
    2. Quantify protein concentration.Note: Quantification of proteins needs to be done with an appropriate method depending on the lysis buffer used, i.e., BCA protein assay (for SDS-containing lysis buffer) and protein assay dye reagent concentrate (lysis buffer without SDS).
    3. Collect the desired quantity of lysates (15–50 µg) in a new Eppendorf, add Laemmli buffer to a final 2× concentration up to 25 µl of volume, and mix well. Then, boil samples at 100 °C for 6 min, let them cool at room temperature and centrifuge at 16,000× gfor 30 s.Note: Samples can be stored at -20 °C if not used immediately.
    4. Load the desired quantity of protein lysates in the corresponding acrylamide gel to perform SDS-PAGE. Place the gel in the tank filled with MES-SDS running buffer and apply current to allow protein separation until the front dye reaches the bottom of the gel ().Note: Depending on the size and proteins to be detected, the percentage and kind of acrylamide gels to be used will be different. For example, to detect proteins with high molecular weight (above 80 kDa), 6%–8% acrylamide gels are preferred. In case it is necessary to detect low (below 30 kDa) and high molecular weight proteins in the same gel, it is recommended to use gradient gels with a percentage of acrylamide from 4% to 12%. Depending on whether the acrylamide gel is a Bis-Tris or Tris-Glycine, the resolving range and the time required to run the samples will vary (30–40 min in Bis-Tris gels at 200 V to 75–120 min in Tris-Glycine gels at 90 V).Figure 1.SDS-PAGE gel running; at loading (left panel) and after running for 30 min (middle panel) and 90 min (right panel).Note the molecular weight standards (blue) appearing on the left and right edges once the samples have run. Arrows point to the front dye and arrowheads to 75 kDa molecular weight marker.
    5. Transfer proteins from gel to PVDF membranes. Before obtaining the gel, label and activate the PVDF membrane by soaking it in methanol for a few seconds, rinse twice with ddH2O for 3–4 min, and incubate with transfer buffer until use. Stop current once the front dye has reached the bottom part of the gel. Assemble the transfer cassette in a container with transfer buffer as shown in.Note: It is crucial to avoid air bubbles when setting the cassette by covering the sponge, Whatman paper, and the gel with transfer buffer. Failure to remove air bubbles in the assembled cassette will jeopardize the detection of proteins. After placing each component, remember to run the blot roller gently to avoid the presence of air bubbles during the transfer. Be careful with the orientation of the gel so as not to lose the order in which the samples were loaded; a different volume of molecular weight marker can be added in the first and last lane to aid this purpose.Video 1.Transfer preparation
    6. Place the cassette in the frame in the right orientation located in the chamber filled with transfer buffer and an ice package. Include a magnetic stirrer and stir while applying a 200–400 mA current for at least 1.5 h in a cold room to avoid overheating. After that, remove PVDF membrane with the proteins attached and start the procedure to detect them.Note: Protein transfer from gel to membrane can be done with different equipment, including semi-dry transfer. Insertion of the cassette in the wrong orientation will make proteins from the gel to migrate in the opposite direction of the PVDF membrane; therefore, proteins will be lost.
    7. Block membrane with blocking solution with agitation for at least 30 min at room temperature.Note: Blocking solution will depend on the primary antibody to be used. Check manufacturer datasheet for that info.
    8. Incubate the membrane with the primary antibody in blocking solution with agitation for 2–3 h at room temperature or overnight at 4 °C.Note: The quality of the signal and the amount of background will depend in a great manner on the quality/affinity of the antibodies. If possible, use antibodies that have been previously tested by you or other laboratories.
    9. Wash with gentle agitation at least three times with TBST solution for at least 15 min, block membrane using TBST + 5% non-fat milk for 10 min with agitation, add the corresponding secondary antibody conjugated with HRP, and incubate for at least 45 min.Note: The secondary antibody to be used can be coupled with different fluorophores instead of HRP, but extra care must be taken with light to avoid bleaching of signal.
    10. Repeat the washes with TBST and leave the membrane in this solution until development.
    11. Set up the settings for image acquisition in the MicroChemi apparatus using different exposure times.Note: Acquisition of images can be performed with other equipment. The exposure times will depend on the quantity of protein and/or selectivity of the primary antibody. Routinely, we set time differences between exposures in the Chemidoc program of 1, 3, 5, 15, 30, 60, 180, 300, 600, and 900 s. The signal of each image corresponds to the accumulated time from the beginning, i.e., images 3 and 4 will have an exposure of 9 and 24 s, respectively.
    12. Incubate membrane with the WesternBrightTMECL HRP Substrate for 1 min.
    13. Place the membrane in the MicroChemi chamber and acquire images ().Note: At the end of the experiment, take a picture with bright light to save an image with the molecular weight markers position in the membrane. This will help identifying the bands of interest.
    14. Copy images and transfer to a computer with ImageJ program to process them.Figure 2.Images obtained with MicroChemi equipment and GelCapture software.Western blots for actin in HEK293FT cell lysates with different exposure times (from 1 to 4,128 s) showing original images (two left panels, the first being inverted), saturation image profile (third panel from the left), and pseudo color image (fourth panel from the left). Note that the signal becomes saturated (red signal) in 37.5 µg sample with 828 s of exposure.
  • Image processing and band quantificationThe aim is to obtain a numerical value for the intensity of each band in each image captured during membrane exposure, using the analysis tools of ImageJ program ().Video 2.Band quantification
    1. Open all images using ImageJ with the commands: File > Import > Image Sequence… choosing the folder directory where the images are located, the first one to start with, and the number of photos to import.Note: Do not include the “Molecular Weight Markers Picture” taken with clear light, only the ones acquired with chemiluminescence.
    2. Choose the parameters of interest to be recorded from the images. To do this, select Analyze > Set Measurements… clicking the ones that ImageJ should record for the analysis.Note: In our case, we have selected Area, Mean gray value, Integrated density, and Display label. The analysis is going to be performed with Integrated density values, but as many parameters as desired can be recorded.
    3. Select an image with visible bands, but not saturated (, overexposure images), and delimit a rectangle that fits one of them with the "Rectangle" tool. Then, move the rectangle over an area of the membrane close to the bands but without signal, to capture the background. Record the measurements with the commands: Image > Stacks > Measure Stack…Note: The less the area of the rectangle fits the band, the greater the possibility of introducing deviations due to the background ([Gassmann et al., 2009]). The size of the rectangle should not be modified, so that the same area is always quantified for all bands. If there are free wells in the gel, a lane with the same lysis buffer and Laemmli buffer can be included, which will serve to quantify the background.
    4. Move the rectangle over each band and repeat the same commands to obtain measurements. As a result, a table with all the values collected for each of the bands in each sequence image will appear.Note: Consider the order in which the bands have been quantified to know to which of the samples the value corresponds.
    5. Save the data in.csv format with the commands File > Save As…
  • Linear model optimizationThe aim of the analysis is to obtain a representative value for each band in the western blot. This value will be the slope of the signal increment within the linear range of the selected band.
    1. Load the required packages for R software to work.Note: If the packages are not installed, the command install.packages() needs to be used before loading them.my_packages <- c("dplyr", "ggplot2", "openxlsx")lapply(my_packages, require, character.only = TRUE)
    2. Set the working directory, the folders where the data file is located, and where the result of the analysis is going to be exported.Note: Remember that quotation marks should be used.WorkingDirectory <- "/Users/xxxxx/Desktop/Folder1/Folder2/Actin"setwd(WorkingDirectory)
    3. Load the data from the image analysis in.csv format, copying the file name in the corresponding lane.Note: Quotation marks should be used.File_Name <- "Results.csv"Results <- read.table(file = File_Name, sep = ",", header = T)
    4. Define the conditions of the experiment: the background and the samples that have been loaded on the gel. The script is going to assign each name to each set of measurements in a new column called “conditions”.Note: Be careful to write them in the same order in which the quantification has been done from the images and use quotation marks so that R interprets them as text.conditions <- c("Background", "6μg", "9μg", "15μg", "18μg", "22.5μg", "30μg", "37.5μg")Results$Conditions <- factor(x = rep(conditions, each = max(Results$Slice)), levels = conditions)
    5. Introduce the exposure time for each image that has been quantified. The script is going to assign each value to its corresponding image in a new column called “Exposure.”Note: Remember that it is not the value that has been introduced in the Chemidoc software, but the accumulated time from the beginning. The units to be used can be seconds or minutes; if using minutes, convert the values to decimal system and use “.” for decimals.Exposure <- c(1, 3, 8, 18, 33, 63, 108, 168, 228, 288, 408, 588, 828, 1128, 1428, 1728, 2028, 2628, 3228, 4128)Results$Exposure <- rep(exposure, times = length(conditions))
    6. Perform background correction. The code will subtract the signal from the background of each image in all analyzed bands. The result will be written in a new column called “BckgCorrect.”Results$BckgCorrect <- rep(0, dim(Results)[1])for(iin1:dim(Results)[1]) {Bkgnd <- Results$Conditions == "Background" &Results$Exposure == Results[i,"Exposure"]Results$BckgCorrect[i] <- Results[i,"IntDen"] -Results[Bkgnd,"IntDen"]}
    7. Represent the data in a scatterplot, with the exposure time as the independent variable and the background corrected signal as the dependent one (). The purpose of the graph is to determine visually the linear range for the increment of the signal.ggplot(Results, aes(x = Exposure, y = BckgCorrect, color = Conditions))+geom_point() + geom_line() + theme_classic()Figure 3.Representation of band quantifications.The signal of each band (y-axis) with respect to the exposure time (x-axis) is shown. Note that in the samples with higher protein content (22.5, 30, and 37.5 µg) the linearity between exposure time and signal is lost at certain exposure time.
    8. Optimize the model by manually selecting the first and last exposure time to be considered that stays within the linear range for all conditions of interest. The script will define a new data frame with the filtered data between these two points, named “Results_filtered.”First_Expossure = 63Last_Expossure = 828Results %>% filter(between(Exposure, First_Expossure, Last_Expossure)) -> Results_filtered
    9. Build the linear model using simple linear regression, with exposure time as predictor variable and corrected signal as response variable. The script will calculate the regression parameters for each condition of the experiment, obtaining a new data frame with the value of the slope of the regression line, the R-squared coefficient of determination to evaluate the goodness of fit, and the p-value to evaluate its statistical significance (). It will also generate a new scatterplot with the filtered data ().Results_model <- data.frame(t(mapply(unlist, by(Results_filtered,Results_filtered$Conditions,function(x)summary(lm(BckgCorrect~Exposure, x)))))) %>%select(slope = coefficients2, adj.r.squared, p.value = coefficients8)ggplot(Results_filtered, aes(x = Exposure, y = BckgCorrect, color = Conditions))+geom_point() + geom_line() + theme_classic()knitr::kable(Results_model)Figure 4.Adjusted representation of band quantifications.After removing the samples with lower and/or higher exposure (in which linearity was lost), linearity for different amounts is kept and will allow to accurately calculate the slope for each concentration.Table 1.Regression parameters for each condition of the experimentConditionsSlopeAdj. r. squaredp. valueBackground0NaNNaN6 μg1157.6330.99888832.704757e-109 μg3156.0740.99939884.277239e-1115 μg16760.630.99869834.34215e-1018 μg26829.650.99940914.061036e-1122.5 μg57338.10.99937254.864208e-1130 μg121563.90.99898342.068317e-1037.5 μg223579.30.99985076.546559e-13
    10. If the regression coefficient is not close enough to 1 and/or the scatterplot does not fit the linear range, go back to step C8, and adjust the exposure values again.
    11. Export the analysis results. The script will overwrite the initial excel file, saving all the data in a sheet named “DATA,” the filtered data between exposure times in a sheet named “DATA_FILTERED,” and the coefficients of the linear model in a sheet named “MODEL.”Note: Remember that these are chemiluminescent signal values. Western blotting is a semiquantitative technique, so the next step in interpreting the results should be to compare these values with a housekeeping or a reference protein (protein of interest/reference protein).write.xlsx(list(DATA = Results, DATA_FILTERED = Results_filtered, MODEL = Results_model), file = "File_Name", row.names = T)
    ExampleAs an example of the method described here, complementary to that published in Sánchez-Sánchez et al. (2023), we show the quantification of TrkA activation after stimulation with nerve growth factor (NGF) for different times. Briefly, we used primary cultures of sensory neurons enriched in TrkA-positive neurons selected with NGF for 5–6 days. We then deprived the culture of neurotrophin for 4 h and stimulated with NGF (50 ng/mL) for the corresponding time. We collected cell lysates and performed western blotting using pTrkA (for activated receptor) (, left panels) and TrkA (, left panels) antibodies, taking the top 140 kDa band as the reference protein. We quantified the signals from the pTrkA and TrkA bands as described in this protocol. We measured the signal intensity of different images for each sample (, exposure blots) and, together with the exposure time settings, defined the combined linear range for each protein (, right plots). As described above, we built a linear regression model. We used the slope as a representative value to calculate the ratio of pTrk to TrkA and plotted the results of independent experiments ().Figure 5.Example of signal quantification.Images with different exposures showing pTrkA (panel A) and TrkA (panel B) signals were quantified in cultured sensory neurons after NGF stimulation for different stimulation times. The colored rectangles in both panels represent the membrane areas used for quantification. A rectangle of another region without signal was utilized as background (not shown in the images). Note that saturation can be detected at higher exposures of the TrkA blot in the pseudo color images (red color). On the right side of each panel there is a graph showing the signal increase curve for each time point. The range between dashed lines indicates the exposures taken to optimize the linear regression model and to perform the quantification within linearity. Note that the saturation in the pseudo color images matches the flattering of the curves in the plots. (C) Quantification of TrkA activation (pTrkA/TrkA ratio). Normalization was performed using the 5 min time point of NGF stimulation. Each colored line represents the results obtained in a single experiment (n = 3). Panels A and B correspond to experiment c in panel C.Useful tips
    1. It is important to include control and experimental/problem samples in the same membrane, to be able to compare since there is always some variability between membranes, even with those carried out in parallel for the whole experiment.
    2. Total protein to be loaded in the gel (usually between 15 and 50 µg) will depend on the abundance of the protein of interest to be detected and the affinity of the antibodies to be used. Loading an excess of total protein can also saturate the membrane during the transfer, and differences between samples could be misleading.
    3. Remember that western blot is a semiquantitative technique and a loading control, housekeeping protein, or reference protein are always needed to normalize the data.
    4. The quality of the signal and the amount of background will depend in a great manner on the quality/affinity of the antibodies. If possible, use antibodies that have been previously tested by you or other laboratories.
    5. When working with R software, do not forget to use quotation marks in what should be text.

Recipes

The following solutions/buffers are obtained using the quantities indicated, and the final concentration of each component is given in brackets.

  1. Lysis buffer2.42 g of Trizma base pH 8.0 (20 mM)8 g of NaCl (137 mM)10 mL of NP-40 (1%)Ultrapure MilliQ water up to 1 LStore at 4 °CJust prior to use, add the corresponding volume to have a final concentration of 1 mM PMSF, 1 μg/mL aprotinin, 2 μg/mL leupeptin, 1 mM sodium orthovanadate, 10 mM NaF, and 20 mM β-glycerophosphate.
  2. 5× Laemmli buffer30.29 g of Trizma base pH 6.8 (250 mM)100 g of SDS (10%)500 mL of glycerol (50%)67 mg of bromophenol blue (0.01%)39.07 g of β-mercaptoethanol (500 mM)Ultrapure MilliQ water up to 1 LStored at -20 °C
  3. Polyacrylamide gelsAcrylamide-bis solution up to the desired percentage, 350 mM Bis-Tris buffer pH 6.4, 0.1% SDS, ultrapure MilliQ water, APS (0.1%), and TEMED (6.6 mM). Mix the reagents in that order and avoid bubbles to prevent patchy polymerization. Prepare 4.5 mL of final volume per resolution gel, and 2 mL with lower acrylamide percentage for the stacking. The use of a thin layer of isopropanol to avoid contact with oxygen during polymerization of the resolving gel also helps to avoid bubbles and to have a clean separation with stacking. Homemade gels can be stored up to 3–4 days at 4 °C in a humid chamber.
  4. Transfer buffer5.8 g of Trizma base pH 8.3 (25 mM)2.9 g of glycine (192 mM)100 mL of methanol (10%)Ultrapure MilliQ water up to 1 L and adjust pHStore at 4 °C and reuse up to five times
  5. TBST washing buffer1.21 g of Trizma base pH 7.5 (10 mM)8.8 g of NaCl (150 mM)0.32 g of EDTA (1 mM)10 mL of Tween-20 (0.1%)Ultrapure MilliQ water up to 1 L and adjust pH
  6. Primary antibody blocking solution100 mL of TBST washing buffer3 g of BSA (3%)32.5 mg of sodium azide (0.05%)Stored at 4 °C
  7. Secondary antibody blocking solution100 mL of TBST washing buffer5 g of non-fat dry milk (5%)Prepare just before use with gentle agitation to avoid foaming

Acknowledgements

We thank Deogracias’ and Arevalo’s lab for input into the manuscript. This work was supported by grant PID2020-113130RB-100 funded by MCIN/AEI/10.13039/501100011033 and by the European Union to J.C.A. D.C.G was granted by Consejería de Educación Junta de Castilla y León and the European Social Fund.